Amino Acid Composition and Terminal Amino Acids of Staphylococcal

Glutamic acid was identified as the N-terminal residue and lysine as the ... onine, tyrosine, and carboxymethylcysteine when present during .... 3 .33...
1 downloads 0 Views 730KB Size
BIOCHEMISTRY

Amino Acid Composition and Terminal Amino Acids of Staphylococcal Enterotoxin B" Leonard Spero, David Stefanye, Peter I. Brecher, Henry M. Jacoby, John E. Dalidowicz, and Edward J. Schantz

ABSTRACT: A complete amino acid analysis is presented for staphylococcal enterotoxin B. The protein is composed solely of amino acids and is particularly rich in aspartic acid and lysine. It has no free sulfhydryl groups and only one disulfide bridge. On the basis of the amino acid composition the molecular weight is 35,380, the isoionic point is 8.70, and the partial specific volume is

S

taphylococcal enterotoxin B is the emetic material elaborated by the S-6 strain of Staphylococcus uureus. A purification procedure permitting the isolation of large amounts of an immunologically and physically homogeneous material is described in an accompanying paper (Schantz, er al., 1965). The toxin is devoid of any lipid, carbohydrate, or nucleic acid. This report presents a complete amino acid analysis of the protein, using the automatic amino acid analyzer, and a comparison with an earlier, largely microbiological analysis (Hibnick and Bergdoll, 1959). It also describes the identification of the amino- (N-) and carboxyl- (C-) terminal amino acids of the toxin, and quantitation of these residues. The implications these analyses have on the structure and conformation of the protein are discussed. Materials and Methods Materials. The enterotoxin was prepared according to the method of Schantz et al. (1965). It was stored as the lyophilized powder in a deep freeze. All the solvents used in chromatography were reagent grade. Anhydrous hydrazine was prepared by distillation under reduced pressure in a nitrogen atmosphere from sodium hydroxide pellets and was stored in a desiccator in foilcovered tubes. Benzaldehyde was freshly distilled before use. Dioxane was distilled over sodium, frozen, and stored in a deep freeze. Phenyl isothiocyanate was purified by distillation in vacuo. Amino Acid Analysis Preparation of Samples. Samples containing approximately 2 mg/ml protein were pipetted into 40-ml Pyrex freeze-drying ampoules. An equal volume of concen~~

1024

* From the U.S. Army Biological Laboratories, Fort Detrick, Frederick, Md. Received December 31, 1964; revised March 15, 1965. Presented in part a t the meeting of the American Society of Biological Chemists in Chicago, Ill., March 1964.

SPERO,

STEFANYE, BRECHER,

JACOBY,

0.731. These values are in excellent agreement with those determined experimentally. Glutamic acid was identified as the N-terminal residue and lysine as the C-terminal residue. Quantitative estimation showed one N-terminal and one C-terminal residue per mole of protein, indicating that the structure is a single polypeptide chain.

trated HCl from a freshly opened bottle was added. The mixture was frozen in a dry ice-acetone bath, evacuated to 50-100 Hg, thawed carefully, and agitated to remove dissolved air. Careful performance of this operation is essential to prevent oxidation of methionine, tyrosine, and carboxymethylcysteine when present during hydrolysis (Crestfield et al., 1963). After being sealed under vacuum the ampoule was wrapped in aluminum foil, weighted to maintain a vertical position, and introduced into a flask of boiling toluene. The temperature of the toluene was maintained with an openair reflux system at 110" + 1 '. After hydrolysis the samples were dried without transfer on a lyophilization apparatus. As the hydrolysates did not remain frozen under conditions usually sufficient for the lyophilization of aqueous solutions, the evaporation process was carried out on the liquid sample. Control experiments indicated that no material was lost by entrainment, and that there was no bumping or boiling of the supercooled acid. The process required from 4 to 6 hours for a 4-ml sample. The crystalline colorless residue was stored in a desiccator over KOH pellets. For analysis, the samples were dissolved in 5.00 ml of 0.2 N sodium citrate buffer at p H 2.2, and 2.00-ml samples were applied to the columns. Calibration of the Analyzer. Protein hydrolysates were analyzed on a Phoenix automatic amino acid analyzer (Model 5000) according to the methods of Spackman et al. (1958). Recently it was reported (Mahowald et al., 1962) that the color yield of the reaction between ninhydrin and an amino acid in the chromatographic eluate decreased progressively but not identically over time for all amino acids, and equipment modifications were recommended to minimize these fluctuations. Consequently, the ninhydrin reservoir in the analyzer was water-cooled by winding a coil about it; a bottle containing Fieser's solution, a potent oxygen scavenger, was interposed between the oil reservoir and the ninhydrin solution. These precautions retarded the rate of

D A L I D O W I C Z , A N D

S C H A N T Z

oxidation of the ninhydrin reagent, as evidenced by the constancy of the integrated chromatographic peak areas of a calibration mixture of amino acids passed through the analytical column before and after the run. As an extra precaution, the same column was used for calibration mixture analyses as well as for the analysis of protein hydrolysates. An average deviation of 1 2 % was observed in the amino acid analyses of calibration mixtures. Calculations.Three separate analyses were performed. The nitrogen content of the protein was uncertain at the time of the first two analyses. Initially, therefore, the weight per cents of the individual amino acid residues, based on the best available dry weight of protein, were summed and the total was divided into 100 to obtain a correction factor. Each of the amino acid percentages was then multiplied by this factor and the amino acid composition of a molecule of 35,000, the molecular weight of the protein as determined by physical methods (Wagman et al., 1965), was calculated. Based on the excellent internal precision in the two runs and the fact that it calculated so closely to a unitary value, glutamic acid was taken as a reference standard. All the values were then recalculated relative to glutamic acid, assuming 26 residues per mole. A similar procedure has been followed by other investigators (Bargetzi et al., 1963). At the time of the last analysis, the per cent nitrogen was known more accurately and was used to determine the size of the sample used in the analysis. Minimal molecular weights were first calculated from the average weight percentages of the individual amino acid residues, and nearest integer numbers for the amino acid residues were obtained for 35,000 g of protein. A new molecular weight was then computed as a first approximation from these numbers and their respective residue molecular weights. These results in turn yielded integral numbers of residues for the recalculated molecular weight, as a second approximation. Summation of the products of these integral numbers multiplied by their respective residue molecular weights yielded a final molecular weight. Further approximations did not change the values for the integral numbers of residues. Other Methods. Cystine was determined (a) as cysteic acid after performic acid oxidation of the protein followed by acid hydrolysis (Moore, 1963); (b) as S-carboxymethylcysteine after reduction with mercaptoethanol, passage through a Sephadex column, alkylation with iodoacetamide, purification on CM-cellulose, and acid hydrolysis; and (c) spectrophotometrically with the reagent developed by Ellman (1959) after reduction with mercaptoethanol and passage through a column of Sephadex G-25. Cysteine was measured by titration with p-mercuribenzoate under a variety of conditions (Boyer, 1954) and as S-carboxymethylcysteine after alkylation in 8 M urea followed by acid hydrolysis. Tryptophan was determined (a) colorimetrically by the method of Spies and Chambers (1949) and (b) by titration with N-bromosuccinimide (Peters, 1959). Amide nitrogen was calculated from the ammonia content of the hydrolysates of varying duration after linear extrapolation to zero time. The threonine and

AMINO

ACID

serine values were calculated in a similar manner. Nitrogen was .determined by micro-Kjeldahl, sulfur by the method of Alicino (1958), and phosphorus by the method of King (1 932). End-Group Analysis Preparation and Hydrolysis of DNP-enterotoxin. The DNP-protein was prepared in a pH-stat (Radiometer TTTlb with recorder SBR2c) at pH 9.0 and 40" according to procedure (3) of Levy (1955). It was hydrolyzed in constant-boiling HCI (5.7 N) in a constant-temperature oil bath at 105" for 16 hours. Ether-soluble DNP-amino acids were extracted after dilution of the hydrolysates to approximately 1 N with respect to the acid. Paper Chromatography of DNP-amino Acids. Both the two-dimensional system of Levy (1955) and the t-amyl alcohol unidimensional descending system of Blackburn and Lowther (1951) were employed. IdentiJication of Terminal PTH-amino Acid.' The paper strip modification of Fraenkel-Conrat (1955) was used for the preparation of the PTH of the N-terminal amino acid. Toxin (25.8 mg) was applied to five paper strips as a 5x solution. The derivative was identified in system F of Edman and Sjoquist (1956) (nheptane, 75 aqueous formic acid, and ethylene chloride, 1 :2 :2) on starch-treated paper. Descending chromatograms were developed for 1.5 hours and the spots were visualized by spraying with iodine-azide. Hydrazinolysis of the Enterotoxin. Samples of toxin solution were introduced into glass ampoules and dried first by lyophilization and then in an oven at 50". Anhydrous hydrazine (0.5 ml) was added in a drybox and the tubes were sealed under vacuum. Hydrazinolysis was carried out at 100" for 10 hours in a constant-temperature oil bath. The solutions were cooled, transferred to evaporating dishes, and dried over H2S04in vacuo. The residue was dissolved in 1.0 ml of water in a small centrifuge tube, 0.25 ml of benzaldehyde was added, and the mixture was shaken for 2 hours. The tube was centrifuged, the aqueous supernatant was withdrawn, and the benzaldehyde treatment was repeated. The combined aqueous layers containing the C-terminal amino acid and some residual hydrazides were then dried. Prior to analysis on the automatic amino acid analyzer the material was dissolved in the appropriate citrate buffer.

x

Results Amino Acid Composition. The elemental composition of the deionized protein was: nitrogen 16.15%; sulfur 1.04%; phosphorus, 0.023%. Table I shows the amino acid composition of staphylococcal enterotoxin B and the calculation of the integral values of the amino acids in the molecule. The results represent the average of three analyses. Each analysis was performed on 24-, 48-, 72-, and 96-hour hydrolysates run singly.

~

~~

Abbreviations used in this work: PTH, phenylthiohydantoin; FDNB, l-fluoro-2,4-dinitrofluorobenzene;PTC, phenylthiocarbamyl-. 1

COMPOSITION

AND

END GROUPS OF

1025

ENTEROTOXIN B

B IOC H E kI ST R

TABLE I : Amino

1

Acid Composition, Integral Residues, and Molecular Weight of Staphylococcal Enterotoxin B.

Amino Acid

Amino Acid Residues woog dry protein),z

Nitrogen (g/100 g protein)

Nearest Integral Number of Amino Acid Residues for 35,380 g Protein

Lysine Histidine Amide ammonia Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Tryptophan Total

15.25 f 0.36. 2.45 f 0.13 1.58d 2.67 0 . 1 2 17.93 f 0 . 3 0 4.69 0 . 0 6 4.23 0.09 9.55 0 . 0 6 2.10 i 0 . 1 8 1 . 9 0 f 0.06 1.37 f 0 . 0 2 0.58 5.49 f 0.14 3.70 0.01 3.45 f 0 . 0 4 6.37 0 . 0 4 11.20i0.34 6 . 1 2 i 0.01 1.05 100.10%

3.33 0.75 1.38 0.96 2.18 0.65 0.68 1.04 0.30 0.47 0.27 0.08 0.78 0.40 0.43 0.79 0.96 0.58 0.16 16.19%

42 6 35d 6 55 16 17 26 8 12 7 2 20 10 11 20 24 15 2 299

* * *

* *

Integral Number of Residues X Respective Residue Molecular Weights

5,384 826 (-34.61, +18.01)e 936 6,331 1,618 1,481 3,357 777 685 498 204 1,982 1,312 1,245 2,264 3,917 2,208 372 35,380

Nearest Integral Number of Residues (lit .)b 47 8 31d 6 59 17 25 25 6 17 6 21 7 12 23 23 14 2 318

a To avoid computational errors due to rounding off, two decimal places are retained throughout the columns. The average deviation is not better than 2%. b Calculated from column 1 of Table I in Hibnick and Bergdoll (1959) for a molecular weight of 35,380. Average deviation. d Omitted from the total. e To correct for the molecular weight difference between OH and NH2, 0.989 is subtracted per amide residue. To correct for the mole of water on the terminal amino acids, the molecular weight of water is added.

1026

The destruction of serine and threonine appeared to follow zero-order kinetics and the extrapolation to zero time was accordingly made from the least-squares solution of these data. The amide value was also calculated from an extrapolation of the least-squares straight line but in this case all the data were combined into one solution because of the greater scatter encountered. The other amino acids were essentially constant during the time of hydrolysis except valine and isoleucine, which appeared to increase after the shortest hydrolysis. The Spies and Chambers (1949) method gave a satisfactory value for tryptophan. It was necessary, however, to apply the enzymatic digestion technique of Harrison and Hofmann (1961) in order to obtain a true tryptophan color with an absorption maximum at 590 mp. Optimum results were attained with a 5-hour digestion period with trypsin-chymotrypsin. The titration with Nbromosuccinimide gave an essentially identical value, but this method has the undesirable feature of requiring the use of a large empirical correction factor. Several

SPERO,

STEFANYE,

BRECHER,

JACOBY,

DALIDOWICZ,

other methods for tryptophan were tried (Duggan and Udenfriend, 1956; Beaven and Holiday, 1952; Noltmann et at., 1962) and the results from all of them indicated 2 residues per molecule. None of the results of these latter methods was satisfactory, however, as an analytical value. From the point of view of reliability, the analytical average deviation of f 2 % introduces an uncertainty into the aspartic acid and lysine residue numbers because these substances are present to the extent of more than 40 residues per molecule. Consequently, they can be accepted as reliable to I 1 residue. In some instances even a much smaller analytical error may lead to the selection of the wrong integral value. An error of but 1.0% in the tyrosine analyses, for example, increases the calculated number of residues from 24.28 to 25.52 and, therefore, the integral number from 24 to 25. Distribution of Sulfur. Table I1 shows the distribution of sulfur-containing amino acids in the protein. The analyses by various methods are consistent with 2 residues of half-cystine and 10 residues of methionine.

AND

SCHANTZ

VOL.

4,

NO.

6,

J U N E

1965

Distribution of Sulfur in Staphylococcal Enterotoxin B. TABLE 11:

Residues per 35,380 g Protein

Method Sulfur (a) From amino acid analysis (b) From elemental analysis Performic acid oxidation (a) Cystine as cysteic acid (b) Methionine as sulfone Alkylation after reduction with mercaptoethanol As S-carboxymethylcysteine Spectrophotometry after reduction with mercaptoethanol As cysteine Reaction with p-mercuribenzoate As free cysteine Direct alkylation with iodoacetate As S-carboxymethylcysteine

12.0 11.5 2 01 8.20

2.06

1.99